AD8601 unity gain buffer

I have applied a signal to the input of the AD8601 which is an exponential function(like discharging and charging waveforms of an RC filter). The signal oscillates between 1.5V and 3.5V @ 800kHz. The AD8601 is in differential configuration and the resistors in the configuration is 100K all. The output is single ended and fed into a comparator. I want output signal to be exactly the same as the input(like a unity gain buffer) but what i get is a triangular signal rather than exponential which has a slightly less frequency e.g.780kHz and slightly have larger peak to peak value and probably has a phase difference(I did not measure the phase difference). I've used plain differential configuration with 4 resistors. AD8601 is fed from 5V supply. What am I doing wrong? Is this part suitable for my purpose? Thanks.

  • 0
    •  Super User 
    on Aug 3, 2013 4:31 PM

         Hello mustafa,

         The AD8601 has a 6V/uS maximum slew rate, which is not fast enough for the signal you're buffering.  The "fast" part of your exponential waveform slews about twice as fast, as shown in blue below.  When the AD8601 encounters such a signal, the best it can do is change its output at the rate of six volts per microsecond, resulting in a straight-line trace (the red line below).

         Note  the AD8601's 8 MHz bandwidth applies to signals small enough to slew less than 6V/uS -- above that level, the slew rate spec takes over.

         Thus for your application you need an op-amp with a slew rating of at least 15V/uS.  An ADI applications engineer would be in a good position to help you find one.  You could help them by thinking about what other specs you could trade-off for the faster slew rate.  For example, do you need ultra-precision DC offset, lowest audio distortion, etc.  The AD8601 offers stellar performance in these areas, making it ideal for many applications.  Its slew rate, however, is just a bit pokey for your purpose.

         Best regards,

         Bob

    Message was edited by: Robert Blowsky -- p-p voltage shown in graph corrected.

    Message was edited to correct the p-p voltage shown in the graph.

  • 0
    •  Analog Employees 
    on Aug 6, 2013 1:09 AM

    Hi mustafa,

    I agree with Bob, it looks like the amplifier is coming up short on slew rate so it can't catch up to your input signal. If you can provide other specifications that are important to your application, please do. Otherwise, I would recommend you take a look at AD8651. It has a slew rate of 41V/us (more than enough). It is rail to rail so it should capture your desired output range, can accommodate a 5V supply and also has low distortion.

    I hope this helps!

    Kris

  • I got it, thanks. Ad8651 seems to be proper. I need to use high resistance value resistors with the opamp and ad8651 seems to be a fit for that too.